🎓 Lesson 5
D3
Calculating Positioning Uncertainty in Multi-Path Canyon Environments
When drones or GNSS receivers try to locate themselves in narrow canyons or between steep rock walls, reflected satellite signals confuse the system—making position estimates less accurate.
🎯 Learning Objectives
- ✓ Calculate 3D positioning uncertainty (σₕ, σᵥ) using GDOP and multipath error models for a given canyon geometry
- ✓ Analyze satellite visibility masks and predict worst-case GDOP using skyplot simulation tools
- ✓ Design antenna placement strategies to minimize multipath in confined mine canyons
- ✓ Explain how carrier-phase ambiguity resolution failure propagates into centimeter-level uncertainty in RTK mode
📖 Why This Matters
In open-pit mines, drone-based surveying often occurs in deep, narrow benches or haul roads flanked by high walls—environments where GNSS signals bounce unpredictably off rock faces. A 10 cm positioning error may seem minor, but when mapping blast-hole collars or verifying stockpile volumes, it introduces systematic bias that compounds across thousands of points. In 2022, a Tier-1 copper mine reported a 7% overestimation in ore volume due to uncorrected canyon-induced GNSS drift—costing $2.3M in unplanned re-surveying. Understanding and quantifying this uncertainty isn’t academic—it’s foundational to survey-grade data integrity.
📘 Core Principles
Multipath error arises when GNSS signals reflect off nearby surfaces before arriving at the antenna, creating constructive/destructive interference that distorts pseudorange and carrier-phase measurements. In canyon environments, the sky view is restricted—reducing the number of visible satellites and worsening GDOP (Geometric Dilution of Precision), which amplifies measurement errors into position errors. Unlike open-sky conditions where errors follow Gaussian distributions, canyon multipath produces skewed, heavy-tailed uncertainty distributions. Modern mitigation combines physical (choke-ring antennas), temporal (signal smoothing), and computational (multipath-aware RTK engines like Trimble Pivot or NovAtel SPAN) approaches—but first, engineers must quantify the baseline uncertainty budget.
📐 Canyon-Adjusted Position Uncertainty Model
The total 3D positioning uncertainty (σ₃D) combines satellite geometry (via GDOP), multipath magnitude (εₘₚ), and receiver noise (εᵣ). For canyon environments, we augment standard RTK uncertainty with a topographic multipath factor (kₘₚ) derived from wall height-to-width ratio (H/W) and material reflectivity (ρ).
💡 Worked Example
Problem: A drone-mounted GNSS receiver operates in a canyon with H = 45 m, W = 30 m (H/W = 1.5), granite walls (ρ ≈ 0.65), GDOP = 5.2, εᵣ = 3 mm (RMS), and εₘₚ₀ = 8 mm (baseline multipath in urban canyon). Calculate σₕ and σᵥ.
1.
Step 1: Compute kₘₚ = 1.0 + 0.4 × (H/W) × ρ = 1.0 + 0.4 × 1.5 × 0.65 = 1.39
2.
Step 2: Estimate multipath error: εₘₚ = kₘₚ × εₘₚ₀ = 1.39 × 8 mm = 11.1 mm
3.
Step 3: Apply uncertainty model: σₕ = GDOPₕ × √(εᵣ² + εₘₚ²) = 3.8 × √(3² + 11.1²) = 3.8 × 11.5 = 43.7 mm; σᵥ = GDOPᵥ × √(εᵣ² + εₘₚ²) = 6.1 × 11.5 = 70.2 mm
Answer:
The result is σₕ ≈ 44 mm and σᵥ ≈ 70 mm—exceeding typical RTK sub-20 mm specifications, confirming need for correction or alternative positioning.
🏗️ Real-World Application
At BHP’s Jimblebar Iron Ore Mine (Pilbara, WA), drone surveys of the 60-m-deep North Pit bench suffered >12 cm horizontal drift during morning hours due to low-elevation satellite reflections off east-facing wall. Engineers deployed a topographic mask analysis using Skydel simulation software, identified poor GDOP (< 8) for satellites below 25° elevation, and installed a ground-plane choke-ring antenna elevated 2.5 m above the drone deck. Post-mitigation, σₕ improved from 112 mm to 28 mm RMS—validated via independent total station tie-in points. This case is documented in the 2023 AusIMM ‘GNSS in Mining’ Best Practice Guide.